Isolation and genomic analysis of otomycosis pathogens, and genomics-informed antimicrobial peptide design with potential therapeutic applications

Abstract

Background

Fungal infections are increasingly recognized as a major public health challenge, and superficial diseases such as otomycosis remain clinically consequential because persistence, recurrence, and treatment failure are common when the causative species and susceptibility profile are not clearly defined. However, otomycosis associated isolates are often insufficiently characterized at the genomic level. Furthermore, the development of antifungal candidates, particularly for short antimicrobial peptides, remains heterogeneous in terms of the integration of efficacy, biosafety, and mechanistic endpoints.

Methods

Here, we combined clinical otoscopy and cerumen culture with whole genome sequencing and multi-database annotation of an otomycosis derived Aspergillus terreus AT-1 isolate, including targeted analyses of cytochrome P450 families, secondary metabolite biosynthetic gene clusters, curated pathogenicity resources, and antimicrobial/antibiotic resistance signatures. Then, we rationally designed five C-terminally amidated, phenylalanine based short peptides with controlled physicochemical parameters and assessed antifungal activity by broth microdilution MIC assays against A. terreus AT-1 and yeast pathogens, complemented by hemolysis testing and propidium iodide based flow cytometry and microscopy to evaluate membrane permeabilization.

Results

Otoscopy was consistent with otomycosis, and culture yielded colonies identified as A. terreus AT-1 from the affected ear but not from recovered controls. Genome annotation defined a functionally diverse repertoire spanning metabolism, transport, secondary metabolism, pathogenicity linked features, and a structured resistance gene landscape. The peptide panel showed measurable antifungal activity with distinct hemolysis profiles, and peptide C3 increased A. terreus AT-1 conidial membrane permeability.

Conclusions

Overall, this study establishes an integrated and reproducible workflow that couples clinical isolation and genome resolved characterization of an otomycosis associated A. terreus AT-1 isolate with biosafety aware, mechanism informed triage of phenylalanine based short peptides.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12866-026-05305-z.

Introduction

Fungal infections are increasingly recognized as a major yet historically under prioritized public health challenge, prompting international calls to strengthen surveillance, diagnostics, and antifungal innovation [1]. Even common superficial infections remain clinically consequential, because persistent symptoms, recurrence, and treatment failure may become more likely when the causative species and its antifungal susceptibility are not clearly defined [2, 3]. Otomycosis (fungal otitis externa) exemplifies this problem in a localized setting, where incomplete organism level information can hinder optimal management and predispose to relapse [3].

Otomycosis is frequent in otorhinolaryngology practice and is often associated with warm or humid environments, prolonged moisture exposure, and disruption of the local epithelial barrier [3]. Culture based studies and meta-analyses indicate that Aspergillus and Candida species account for most cases, but the contribution of specific species varies across regions and patient populations, underscoring the need for organism level diagnosis rather than syndrome based treatment [3]. Within Aspergillus, non-fumigatus species are increasingly important because they can show distinct susceptibility profiles and persistence traits that affect therapeutic choice. Among them, A. terreus is clinically notable: reviews and comparative susceptibility analyses repeatedly report reduced susceptibility to amphotericin B, complicating empiric therapy when species identification is delayed [2]. These challenges are compounded by broader constraints, including a limited array of drug classes, inconsistent access to timely diagnostics, and mounting concerns regarding resistance. This underscores the necessity for antifungal strategies that are mechanistically diverse [4].

In this context, genome enabled profiling offers a high resolution workflow to contextualize otomycosis associated isolates and generate mechanistic hypotheses that complement phenotypic susceptibility testing. Draft genomes can define repertoires linked to metabolism, transport, stress tolerance, and potential host interaction, and these can be interpreted using curated resources that connect genes with pathogenicity phenotypes, virulence factors, and antimicrobial/antibiotic resistance. In fungi, this approach proves to be especially informative. This is due to the fact that secondary metabolism and detoxification programs, which include biosynthetic gene clusters (BGCs) and cytochrome P450 families, may influence ecological fitness, chemical tolerance and antifungal response [5, 6]. Consistent with this perspective, resistance associated genes in the A. terreus genome span multiple categories, including efflux/transport systems, sterol and membrane lipid pathways, detoxification and drug metabolism, antifolate related processes, and drug modifying enzymes. This distribution implies that intrinsic resistance in A. terreus reflects a multifactorial network rather than reliance on a single target or pathway.

Given such a heterogeneous resistance landscape, antimicrobial peptides (AMPs) represent a complementary antifungal modality. Strategies driven by membrane directed physicochemical interactions may offer advantages over small molecule drugs that inhibit single intracellular enzymes. AMPs are typically short, cationic molecules whose activity can be tuned by sequence encoded properties (e.g., charge, hydrophobicity, and aromaticity), enabling rational exploration of potency and selectivity. Many antifungal AMPs act through envelope directed mechanisms, including membrane perturbation, and peptide engineering can further refine performance [7]. Because membrane active peptides do not depend on a single intracellular enzymatic target, they may be less susceptible to rapid target site mutation driven resistance [7]. However, translation requires balancing antifungal efficacy with host compatibility, since increased hydrophobicity and amphiphilicity can elevate hemolysis or other cytotoxicity [7]. Furthermore, the presence of multiple transporter families, particularly ABC(ATP-binding cassette) and MFS(Major facilitator superfamily)/PDR(Pandrug resistance system) systems, underscores the necessity to consider efflux susceptibility in conjunction with membrane interaction properties [2]. Optimizing hydrophobicity, amphipathicity, and charge distribution may facilitate the reduction of transporter mediated clearance while reducing dependence on a single intracellular target [7, 8].

Despite progress in clinical mycology and peptide engineering, two practical gaps remain in otomycosis relevant settings. First, otomycosis associated isolates are often managed with limited molecular characterization, leaving isolate specific genomic features relevant to persistence, stress tolerance, and treatment response insufficiently defined [3]. This is particularly important for Aspergillus species with atypical profiles, where genome scale annotation can improve mechanistic hypotheses and experiments by revealing secondary metabolite potential, curated virulence/host interaction mappings, and antimicrobial/antibiotic resistance signatures. Second, short peptide design studies vary in how they integrate spectrum testing, biosafety readouts (e.g., hemolysis), and cellular level mechanistic probes that connect growth inhibition to interpretable endpoints, limiting comparability and slowing translation [7].

To address these gaps, we coupled genome resolved characterization of an otomycosis associated A. terreus AT-1 isolate with the rational design and functional evaluation of short phenylalanine (F) based peptides. We (i) isolated and identified the putative pathogen from clinical cerumen and generated a draft genome; (ii) assembled the genome, predicted key components, and performed multi-database functional annotation with targeted analyses of cytochrome P450 families, secondary metabolite BGCs, curated pathogenicity/drug resistance resources (PHI-base and DFVF), and antimicrobial/antibiotic resistance; and (iii) designed a small panel of thirteen amino acid, C-terminally amidated F-based peptides with controlled physicochemical parameters and evaluated antifungal activity, hemolysis as a biosafety proxy, and membrane permeabilization using propidium iodide (PI) based readouts. We hypothesized that integrating isolate centered genomics with constrained peptide design would link pathogen context to peptide performance and facilitate prioritization of candidates with a balanced antifungal and biosafety profile.

The following sections describe clinical isolation and cultivation procedures, the genome sequencing and annotation workflow, and peptide design and experimental assays, followed by genomic results and antifungal, hemolysis, and membrane permeabilization evaluations, and a discussion of implications for otomycosis relevant management and antifungal peptide development.

Methods

All fungal experiments were conducted under Biosafety Level 2 (BSL-2) conditions in the P2 laboratory at the School of Public Health (Shenzhen), Sun Yat-sen University, China.

The isolation of pathogenic fungi

Isolation of pathogenic fungi required the establishment of both experimental and control groups. Cerumen samples for the experimental group were collected from the affected ears of one patient with otomycosis, whereas control samples were obtained from five ears that had recovered after treatment at the Department of Otorhinolaryngology, the Fifth Affiliated Hospital of Sun Yat-sen University in Zhuhai, China. The collected cerumen samples were mixed with phosphate buffered saline (PBS) and serially diluted to 10^− 5. Subsequently, 200 µL of each dilution was evenly spread onto YPD medium and incubated at 37 °C for 7–14 days to assess fungal growth. Next, the isolated fungus was subjected to whole genome sequencing and genome assembly, followed by genome component prediction and gene function analysis to comprehensively characterize its features related to metabolism, transport, pathogenicity, and antimicrobial/antibiotic resistance. The fungal strain isolated in this study were named as A. terreus AT-1.

Genome sequencing

Genomic DNA was extracted using the SDS method [9]. The extracted DNA was assessed by agarose gel electrophoresis and quantified using a Qubit^® 2.0 Fluorometer (Thermo Scientific). For library preparation, 1 µg of DNA per sample was used as input. Sequencing libraries were constructed using the NEBNext^® Ultra™ DNA Library Prep Kit for Illumina (NEB, USA) according to the manufacturer’s instructions, and index codes were added to enable sample identification. Briefly, genomic DNA was fragmented by sonication to an average size of approximately 350 bp; the fragments were then end repaired, A tailed, and ligated to full length Illumina adapters, followed by PCR amplification. Finally, PCR products were purified using the AMPure XP system, and libraries were evaluated for size distribution using an Agilent 2100 Bioanalyzer and quantified by real time PCR. Whole genome sequencing of A. terreus AT-1 was performed on the Illumina NovaSeq platform (PE150) at Beijing Novogene Bioinformatics Technology Co., Ltd.

Genome assembly

The raw sequencing data (raw data) contained a proportion of low quality reads. To ensure the accuracy and reliability of downstream analyses, the raw data were filtered to generate clean data (clean data). The filtering criteria were as follows: (1) reads with low quality bases (Q ≤ 20) exceeding a threshold proportion (default: 40%) were removed; (2) reads with ambiguous bases (N) exceeding a threshold proportion (default: 10%) were removed; (3) reads with adapter contamination, defined as an overlap with the adapter sequence exceeding a threshold length (default: 15 bp) with fewer than three mismatches, were removed; and (4) if host contamination was suspected (e.g., for minigenome samples), reads were aligned against the host reference using BLAST to remove host derived sequences.

Clean data were then used for genome assembly. First, assemblies were generated using SOAPdenovo with different k-mer sizes (default: 95, 107, and 119), and the optimal assembly was selected based on project requirements using the k-mer that produced the fewest scaffolds, followed by parameter tuning (e.g., -d, -u, -R, and -F) to obtain a preliminary assembly. Second, gaps in the preliminary assembly were filled using GapCloser, and potential lane-specific contamination was reduced by removing reads with low sequencing depth (< 0.35× the average depth) to obtain the final assembly. Finally, contigs shorter than 500 bp were discarded, and the resulting assembly was used for gene prediction.

Genome component prediction

Genome component prediction included the identification of protein-coding genes, repetitive sequences, and non-coding RNAs. The analyses were performed as follows. For fungi, protein coding genes were predicted using AUGUSTUS (v2.7). Interspersed repeats were identified using RepeatMasker. Tandem repeats were detected using Tandem Repeats Finder (TRF). Transfer RNA (tRNA) genes were predicted with tRNAscan-SE. Ribosomal RNA (rRNA) genes were identified using RNAmmer.

Gene function analysis

Predicted protein sequences were annotated against seven databases, including GO (Gene Ontology) [10], KEGG (Kyoto Encyclopedia of Genes and Genomes) [11, 12], KOG (Clusters of Orthologous Groups) [13], NR (Non-Redundant Protein Database) [14], TCDB (Transporter Classification Database) [15], CAZy database (The Carbohydrate-Active EnZymes database), and Swiss-Prot [16], using BLASTp searches with an E-value cutoff of 1e − 5, minimum alignment coverage of 40%, and minimum sequence identity threshold of 40%. Cytochrome P450 genes and secondary metabolite gene clusters were annotated using the P450 database and antiSMASH, respectively. To assess pathogenicity related genes, PHI-base (Pathogen–Host Interactions database) [17] and DFVF (Database of Fungal Virulence Factors) were used.

To profile the putative antimicrobial/antibiotic resistance-associated features of the A. terreus AT-1 genome, we integrated annotation outputs from NR, Swiss-Prot, Pfam, KEGG, TCDB, and PHI-base databases and screened gene records for resistance associated signatures (e.g., multidrug/efflux terms, drug/antibiotic resistance keywords, xenobiotic/drug metabolism functions, and resistance linked KEGG pathways). Candidate genes were then grouped into mechanistic categories (efflux/transport, sterol/membrane-lipid targets, detoxification/drug metabolism, antifolate linked resistance, drug modifying enzymes, and other/unspecified), and membrane associated resistance was defined a priori as efflux/transport determinants together with sterol/ergosterol related targets. Finally, for membrane associated transporter genes, TCDB assignments (TC numbers and descriptions) were used to further stratify efflux/transport candidates into ABC, MFS/PDR, and other transport subclasses.

Synthesis of F-based peptides

Five phenylalanine based peptides (C1, C2, C3, C4, and C5) were synthesized by GL Biochem Ltd. (Shanghai, China) using 9-fluorenylmethyloxycarbonyl (Fmoc)-based solid-phase peptide synthesis (SPPS). The crude peptides were purified by RP-HPLC to a final purity of > 95%, and their molecular weights were confirmed by MALDI-TOF MS (Linear Scientific Inc., USA). All peptides were C-terminally amidated to enhance chemical stability and resistance to proteolytic degradation.

Fungal strains and cultivation

This study included A. terreus AT-1, Candida albicans ATCC 10,231, Candida krusei ATCC 6258, and Candida glabrata ATCC 90,030, all of which were preserved in our laboratory (Shenzhen Key Laboratory of Systems Medicine for Inflammatory Diseases, School of Medicine, Shenzhen Campus of Sun Yat-Sen University, Shenzhen, PR China). Fungi were cultured aerobically in YPD medium at 37 °C. Fungal suspensions at the logarithmic growth phase were harvested for subsequent analyses.

Antifungal activity by MIC assay

The antimicrobial activity of the peptides was evaluated by determining the MIC in 96 well plates using a slightly modified broth microdilution method. Fungal strains were cultured to an OD600 of 0.4, pelleted, washed, and resuspended in YPD to a final concentration of 1 × 10⁵ CFU/mL. YPD was used to maintain consistent growth across filamentous and yeast fungi under our assay conditions. In parallel, peptides (C1, C2, C3, C4, and C5) were serially two fold diluted in PBS in 96 well plates. An equal volume of the fungal suspension was then added to each well. Plates were incubated at 37 °C for 24–36 h, and the MIC was defined as the lowest peptide concentration that completely inhibited visible fungal growth. The tested fungi included A. terreus AT-1, C. albicans ATCC 10,231, C. krusei ATCC 6258, and C. glabrata ATCC 90,030. Fluconazole and wells containing untreated fungal cultures and medium served as positive and negative controls, respectively. All experiments were performed in triplicate. In addition, we used the same approach to assess the antifungal activity of peptide C3 against A. terreus AT-1 conidia.

Hemolysis assay

The hemolytic activity (early stage screening) of peptides C1, C2, C3, C4, and C5 was evaluated using freshly isolated murine red blood cells from one C57BL/6 mice. Specifically, murine blood was collected via the submandibular vein, with a collection volume of 0.2–0.5 mL per mouse. The procedure was performed under strict aseptic conditions to minimize hemolysis and contamination. After blood collection, mice received warmed isotonic fluid supplementation and were monitored to ensure recovery. At the end of the experiment, mice were first anesthetized with 2% isoflurane. While under anesthesia, mice were euthanized by cervical dislocation. All anesthesia and euthanasia procedures were performed in accordance with standardized safety protocols to minimize unnecessary suffering.

Erythrocytes were collected by centrifugation (1,000 × g, 10 min), washed with PBS, and diluted in PBS to 10% (v/v). The erythrocyte suspension was then mixed with an equal volume of serially diluted peptides in a 96 well plate to achieve final peptide concentrations of 32–128 µg/mL, followed by incubation at 37 °C for 2 h. After centrifugation, supernatants were transferred to a new 96 well plate, and hemoglobin release was quantified by measuring absorbance at 570 nm. PBS treated erythrocytes and 0.1% Triton X-100 treated erythrocytes served as negative and positive controls, respectively. Each condition was tested in triplicate, and at least three independent experiments were performed. The hemolysis assay protocol was adapted from previously published antimicrobial peptide hemolysis methods with minor modifications [18].

Cell membrane permeability analysis

Cell membrane permeability was assessed with slight modifications to previously described methods [19, 20]. PI enters cells with compromised membranes and binds nucleic acids, thereby enabling the identification of membrane permeabilized (nonviable) cells. To evaluate the effects of peptide C3 on A. terreus AT-1, 500 µL of conidial suspension (1 × 10⁶ conidia/mL) was treated with C3 at 0 and 16 µg/mL and incubated at 37 °C for 3 h. PI was then added to a final concentration of 20 µM, and samples were incubated at 37 °C for 1 min or 30 min. Fluorescence intensity was quantified by flow cytometry (NovoCyte Advanteon V6B5R3, Agilent, China). Data were analyzed using NovoExpress software.

To more comprehensively assess the membrane disruption dynamics of peptide C3 in A. terreus AT-1 conidia, conidia were harvested by centrifugation 3 h after C3 treatment, resuspended in Tween-80, dual stained with Hoechst and propidium iodide (PI) for 15 min, washed twice with PBS, and then visualized using an inverted fluorescence microscope (Eclipse Ti2-E, Nikon, Japan).

Statistical analysis

Quantitative data are presented as mean ± standard error of the mean (SEM). Significance analysis was performed by independent sample t-test between two groups or one-way ANOVA among multiple groups with GraphPad Prism version 9.3 (San Diego California, USA), followed by Tukey’s post hoc test. A difference was deemed statistically significant at *P < 0.05, **P < 0.01, or ***P < 0.001. Ns, not significant, P > 0.05.

Results

Genome assembly and multi-database annotation of an otomycosis derived A. terreus AT-1 isolate

Otoscopy revealed abundant gray-white/yellow-white flocculent-to-plaque-like adherent material and debris-like discharge within the external auditory canal, with focal clumped accumulations coating the canal wall and partially obscuring visualization (Fig. 1A). The surrounding canal skin appeared hyperemic and swollen, consistent with a marked inflammatory response (Fig. 1A). Overall, these findings are consistent with typical otoscopic features of otomycosis, characterized by fungal debris and mycelial clump-like deposits accompanied by inflammation of the external auditory canal mucosa (Fig. 1A). However, the causative fungal pathogen cannot be reliably determined based on otoscopic appearance alone. Then, cerumen from the affected ear of one patient was cultured in YPD medium, yielding fungal colonies identified as A. terreus AT-1 (Fig. 1A and C and Table S6). In contrast, no comparable fungal colonies were recovered from cerumen collected from the recovered ears of five patients. Collectively, these findings indicate that A. terreus AT-1 was the predominant fungal isolate recovered from the diseased ear and may represent a major contributor to the observed otomycosis phenotype in this patient.

Fig. 1.

Genome assembly and multi-database annotation of an otomycosis-derived A. terreus AT-1 isolate. A Otoscopic examination of a patient with otomycosis and colony morphology of A. terreus AT-1 on YPD medium; B Gene length distribution in the A. terreus AT-1 genome; C Numbers of genes matched to the top 10 species in the NR annotation of the A. terreus AT-1 genome; D KEGG pathway annotation of the A. terreus AT-1 genome; E P450 classification of the A. terreus AT-1 genome

To characterize this clinical isolate at the genomic level, we performed short-read whole genome sequencing followed by standard bioinformatic analyses, including genome assembly and functional annotation. For A. terreus AT-1 genome sequencing, a 350 bp insert size library with 150 bp paired-end reads generated 4,623 Mb of raw data; after quality filtering (13.91% reads removed), 3,980 Mb of clean data were retained with a GC content of 50.55% and high base quality (Q20: 96.99%; Q30: 91.7%) (Table S1). The clean data were then processed using a fungal genome analysis pipeline that encompassed genome assembly, genome component prediction (repetitive sequences, protein coding genes, and non coding RNAs), and comprehensive functional annotation against multiple databases (Pfam, KOG, KEGG, TCDB, CAZy, NR, and GO). Additional downstream analyses included cytochrome P450 classification, secondary metabolite BGC prediction, and pathogenicity/drug-resistance annotation using PHI-base and DFVF (Fig. S1).

The final assembly comprised 910 scaffolds (> 500 bp) and 1,165 contigs, consistent with a draft genome generated using short-read Illumina sequencing, with total lengths of 38,309,077 bp and 38,306,432 bp, respectively, and an overall GC content of 50.84% (Table S2). Comparative genomic analysis against publicly available Aspergillus genomes demonstrated that genome size, predicted gene number, and GC content of A. terreus AT-1 fall within the expected range for the genus (Table S3), supporting the biological plausibility and taxonomic consistency of the assembly. The assembly therefore represents a high quality draft genome generated using short read sequencing data. Gene prediction identified 7,827 genes spanning a total length of 10,439,646 bp (27.25% of the genome), with an average gene length of 1,334 bp (Table 1). The predicted gene number was lower than that reported for the reference strain A. terreus NIH2624 (Table S3). This difference may reflect variation in genome assembly and annotation workflows, including differences in gene prediction algorithms and parameter settings. In addition, strain-specific genomic diversity among A. terreus isolates may contribute to variation in gene content. Consistent with these results, the gene length distribution showed that most predicted genes clustered between 200 and 2,000 bp, peaking at approximately 1,000–1,100 bp (Fig. 1B). The genome wide GC content was 58.31%, whereas internal (intergenic) regions accounted for 27,869,431 bp (72.75% of the genome) with an internal GC content of 48.04% (Table 1).

Table 1.

A.terreus AT-1 gene statistics

Genome Size:	38,309,077
Gene Number:	7,827
Gene Length:	10,439,646
GC Content:	58.31
% of Genome(Genes):	27.25
Gene Average Length:	1,334
Gene Internal Length:	27,869,431
Gene Internal GC Content:	48.04
% of Genome(internal):	72.75

In the A. terreus AT-1 genome, 226 tRNA genes (average length 84 bp; total length 19,005 bp; 0.0496% of the genome) and 34 predicted 5 S rRNA genes (average length 115 bp; total length 3,919 bp; 0.0102% of the genome) were identified, whereas no 5.8 S, 18 S, or 28 S rRNA genes were detected in this annotation set (Table S4). Repetitive element annotation using RepeatMasker/Repbase identified 1,834 interspersed repeat elements spanning 483,768 bp (1.2628% of the genome) with an average length of 268 bp. These repeats comprised LTRs (834; 209,038 bp; 0.5457%; 253 bp), DNA transposons (617; 195,773 bp; 0.511%; 320 bp), LINEs (267; 75,039 bp; 0.1959%; 283 bp), SINEs (88; 5,622 bp; 0.0147%; 66 bp), RC elements (14; 1,338 bp; 0.0035%; 96 bp), and unknown repeats (14; 948 bp; 0.0025%; 68 bp) (Table S5). Tandem repeat analysis further identified 2,492 tandem repeats (repeat size 1–657 bp) spanning 154,218 bp (0.4026% of the genome), including 1,693 minisatellite DNA repeats (10–60 bp; 74,285 bp; 0.1939%) and 308 microsatellite DNA repeats (2–6 bp; 12,619 bp; 0.0329%) (Table S6).

Following assembly and genome component prediction, we next performed functional annotation to contextualize the predicted gene repertoire. NR annotation indicated that A. terreus AT-1 was the predominant match among the top 10 species, whereas the remaining matched genes mainly aligned to closely related Aspergillus species, with a small fraction mapping to other fungal genera (e.g., Rhizopus, Penicillium, and Linnemannia) (Fig. 1C and Table S7). KEGG pathway annotation showed that predicted genes were assigned to all five major KEGG functional categories, with the largest proportion associated with metabolism, followed by genetic information processing, and further assignments to cellular processes, environmental information processing, and human diseases (Fig. 1D and Table S7). Consistent with these functional trends, KOG classification revealed broad distribution across multiple functional classes, with notable enrichment in general function prediction only, posttranslational modification/protein turnover/chaperones, and translation/ribosomal structure and biogenesis, together with substantial representation in amino acid transport and metabolism and carbohydrate transport and metabolism (Fig. S2 and Table S7). CAZy annotation further identified genes spanning six enzyme classes, dominated by glycoside hydrolases and followed by glycosyl transferases, carbohydrate binding modules, auxiliary activities, carbohydrate esterases, and polysaccharide lyases (Fig. 1E and Table S7). Transporter Classification Database (TCDB) analysis indicated that transporter related genes were primarily assigned to electrochemical potential driven transporters, followed by primary active transporters and channels/pores, with additional assignments to accessory factors involved in transport, transmembrane electron carriers, group translocators, and incompletely characterized transport systems (Fig. S3 and Table S7). Finally, the Gene Ontology (GO) standard is employed to assign genes across the three primary GO domains: molecular function, cellular component, and biological process. The majority of annotations are concentrated in catalytic activity and binding (molecular function), cell part/cell and organelle-related terms (cellular component), and metabolic process and cellular process (biological process) (Fig. S4 and Table S7).

The metabolic, transport, pathogenicity and antimicrobial/antibiotic resistance features of A. terreus AT-1 genome

To further interrogate features relevant to secondary metabolism, pathogenicity and antimicrobial/antibiotic resistance, we conducted specialized annotations. P450 classification of the A. terreus AT-1 genome showed that cytochrome P450 related genes were distributed across multiple classes, with the highest representation in the P450 E-class group I category and additional assignments to P450 E-class group IV, CYP52, CYP2D, and pisatin demethylase-like groups; a subset of sequences remained undetermined (Fig. 2A and Table S7 and S10). Secondary metabolite biosynthetic gene cluster (BGC) analysis further identified diverse BGC types in the A. terreus AT-1 genome, dominated by nonribosomal peptide synthetase (NRPS), type I polyketide synthase (T1PKS), and terpene clusters, with additional clusters including NRPS-like, indole, and betalactone classes (Fig. 2B). DFVF disease host-key classification of the A. terreus AT-1 genome showed that virulence factor associated matched genes were distributed across multiple host categories, with the highest representation in Vertebrata, followed by Plants, Animals, Xyloid, and Herb, indicating broad host associated virulence annotations (Fig. 3A and Table S7 and S9). DFVF disease-key classification further assigned these virulence associated matches to a wide range of disease categories, with the largest numbers linked to invasive candidal disease and several plant disease types (e.g., leaf spot, blight, and fruit rot), supporting extensive DFVF related pathogenicity annotations (Fig. 3B and Table S7 and S9). PHI phenotype classification demonstrated that PHI-base matched genes spanned multiple phenotype categories, with most annotations falling into reduced virulence and unaffected pathogenicity, followed by loss of pathogenicity; smaller subsets were associated with lethal, increased virulence (hypervirulence), chemistry target: resistance to chemical, and effector (plant avirulence determinant) phenotypes, while a minor fraction was labeled as not available (NA) (Fig. 3C and Table S7 and S8).

Fig. 2.

The metabolic and transport features of A. terreus AT-1 genome. A Analysis of secondary metabolite gene clusters in the A. terreus AT-1 genome; B CAZy class annotation of the A. terreus AT-1 genome

Fig. 3.

The pathogenicity features of A. terreus AT-1 genome. A DFVF disease host-key classification of the A. terreus AT-1 genome; B DFVF disease-key analysis of the A. terreus AT-1 genome; C PHI mutant phenotype classification of the A. terreus AT-1 genome

Across the integrated annotation set, we identified 291 putative antimicrobial/antibiotic resistance associated genes in the A. terreus AT-1 genome (Table S11). Mechanism-based categorization further indicated that non–membrane associated functions predominated (190/291, 65.3%), driven mainly by detoxification/drug metabolism (139 genes, 47.8%) (Fig. 4A-B and Table S11). In contrast, membrane associated resistance determinants accounted for 101/291 genes (34.7%) and comprised efflux/transport candidates (70 genes, 24.1%) together with sterol/membrane lipid associated targets (31 genes, 10.7%) (Fig. 4A-B and Table S11). Within the membrane associated efflux/transport subset (n = 70), TCDB based stratification showed that other transport systems represented the largest fraction (34/70, 48.6%), followed by ABC transporters (19/70, 27.1%) and MFS/PDR type transporters (17/70, 24.3%) (Fig. 4C and Table S11). KEGG mapping of this candidate set highlighted detoxification and metabolism linked pathways, including glutathione metabolism (30 genes), metabolism of xenobiotics by cytochrome P450 (27 genes), drug metabolism cytochrome P450 (20 genes), and steroid biosynthesis (20 genes), with smaller contributions from antifolate resistance (8 genes), ABC transporters (3 genes), β-lactam resistance (3 genes), and antimicrobial drug resistance modules (3 genes) (Fig. 4D and Table S11).

Fig. 4.

Genome-encoded antimicrobial/antibiotic resistance landscape of A. terreus AT-1. A Proportion of membrane-associated versus non–membrane-associated resistance-associated genes (n = 291); B Mechanism-based category composition of resistance-associated genes (n = 291); C Subclassification of membrane-associated efflux/transport candidates (n = 70) into ABC, MFS/PDR, and other transport systems; D KEGG pathways represented among resistance-associated candidates (n = 291)

Designing short F-based peptides for antifungal exploration: fixed charge and hydrophobicity with an aromatic motif core

Building on the genomic and functional annotation of A. terreus AT-1, which highlighted multiple transporter associated resistance determinants and membrane related functional features, we sought to design antimicrobial peptides that primarily target fungal membranes through physicochemical interactions while minimizing susceptibility to transporter mediated clearance. In particular, the presence of ABC and MFS/PDR transporter related genes in the genome suggests that antifungal activity may be compromised by efflux mediated clearance, thereby motivating a design strategy focused on enhancing membrane interaction and physicochemical stability rather than intracellular target engagement.

Guided by these genome-informed functional annotations, we developed a rational peptide design strategy aimed at generating phenylalanine (F)-based peptides with tunable physicochemical properties. Phenylalanine rich sequences were selected based on their well-documented ability to form aromatic hydrophobic cores that promote membrane interaction and self-assembly, which is advantageous for membrane-active antifungal peptide design. Based on this strategy, five peptides were designed and synthesized (Fig. 5A–E; Table 2): C1, C2, C3, C4, and C5. The design incorporated three guiding principles.

Fig. 5.

Schematic structures of five short peptides. A–E represent peptides C1, C2, C3, C4, and C5, respectively

Table 2.

Sequences of F-based peptides and their key physicochemical parameters

Peptide name	Sequence	Percentage of hydrophobic amino acid	Net charge	Number of amino acid residues
C1	KIRVLLSAIKKFF	53.85%	+ 4	13
C2	KILRVVYKILSKF	53.85%	+ 4	13
C3	RAFRVLLSLLRKF	53.85%	+ 4	13
C4	SKILKLFRRLLKF	53.85%	+ 4	13
C5	KVLALLRKIFASF	53.85%	+ 4	13

First, the peptide length was restricted to thirteen amino acids, which helps reduce synthesis cost while limiting potential toxicity toward mammalian cells.

Second, the proportion of hydrophobic residues (53.85%) was maintained to preserve sufficient membrane affinity, enabling stable interaction with fungal membranes while avoiding excessive hydrophobicity that could increase nonspecific toxicity.

Third, the net positive charge was fixed at + 4, ensuring strong electrostatic attraction to negatively charged fungal membranes, a broadly applicable mechanism for membrane active antimicrobial peptides.

Short F-based peptides exhibit broad antifungal activity with distinct hemolysis profiles

MIC assays showed that C1–C5 exhibited antifungal activity against the vegetative cells of A. terreus AT-1, Candida albicans ATCC 10,231, Candida krusei ATCC 6258, and Candida glabrata ATCC 90,030, with C3, C4, and C5 showing the strongest inhibition against A. terreus AT-1 among the tested peptides; fluconazole (FCZ) served as the reference control (Fig. 6A). Hemolysis assays using murine red blood cells further indicated low hemolytic activity for C1–C4 across the tested concentrations, whereas C5 induced markedly higher hemolysis in a concentration-dependent manner (Fig. 6B). Based on integrated screening by MIC and hemolysis assays, C3 was selected as the preferred candidate for subsequent membrane-permeabilization/disruption studies because it retained strong inhibitory activity against A. terreus AT-1 while exhibiting a more favorable biosafety profile (lower hemolysis than C4 and markedly lower than C5).

Fig. 6.

Antifungal activity and hemolytic activity of five short peptides. A MICs of the five short peptides against the vegetative cells of pathogenic fungi A. terreus AT-1, C. albicans ATCC 10,231, C. krusei ATCC 6258, and C. glabrata ATCC 90,030; B Hemolytic activity of the five short peptides against murine red blood cells. In (A), according to the default color-mapping scheme, warmer color blocks indicate a higher degree of fungal inhibition, whereas cooler color blocks indicate a lower degree of fungal inhibition

Analysis of cell membrane permeability of A. terreus AT-1 conidia after peptide C3 treatment

After treating otomycosis, residual A. terreus AT-1 conidia may contribute to relapse and therapeutic failure. Therefore, evaluating how peptide C3 affects conidial membrane permeability in A. terreus AT-1 has translational relevance for downstream applications. Consistent with this rationale, C3 showed a MIC of 16 µg/mL against A. terreus AT-1 conidia, which was higher than that against vegetative cells (8 µg/mL) but remained within the biosafety range according to hemolysis testing (Figs. 6A-B and 7A). To better define the mechanism underlying C3 activity against A. terreus AT-1 conidia, we assessed membrane permeability after C3 exposure. Flow cytometry showed significantly increased propidium iodide (PI) fluorescence in conidia treated with 16 µg/mL C3 for 3 h compared with untreated controls (Fig. 7B-C, P < 0.001). Hoechst/PI co-staining further supported membrane compromise: untreated conidia germinated normally into mycelia with Hoechst-positive nuclei, whereas C3 treatment increased PI uptake, indicating impaired membrane integrity (Fig. 7D). Collectively, these data demonstrate that C3 increases A. terreus AT-1 conidia membrane permeability, leading to cellular damage and enhanced PI entry.

Fig. 7.

Analysis of cell membrane permeability of A. terreus AT-1 conidia after peptide C3 treatment. A The MIC of C3 on A. terreus AT-1 conidia; B Fluorescence intensity of PI staining strain treated with C3 (16 µg/mL) by a flow cytometer; C PI staining positive rate of A. terreus AT-1 conidia after C3 (16 µg/mL) treatment; D Fluorescent staining of Hoechst/PI on A. terreus AT-1 after C3 (16 µg/mL) treatment by a fluorescence microscope

Discussion

In the present study, we combined genome resolved characterization of an otomycosis derived A. terreus AT-1 isolate with the rational design and functional evaluation of short phenylalanine based peptides, thereby establishing a context informed workflow for antifungal candidate prioritization. The genomic analyses delineated a broad functional repertoire linked to metabolism, transport, cytochrome P450-associated functions, secondary metabolite biosynthesis, curated pathogenicity related resources and antimicrobial/antibiotic resistance. Then, the peptide analyses identified C3 as the most balanced candidate because it combined antifungal activity with low hemolytic liability and measurable membrane permeabilizing effects on conidia. Together, these findings suggest that isolate centered genomics can provide a useful biological backdrop for therapeutic exploration and that short membrane active peptide scaffolds may represent a complementary strategy for otomycosis relevant fungal control.

A major inference from the genomic data is that this otomycosis derived A. terreus AT-1 isolate possesses a functionally rich genetic landscape compatible with ecological adaptability and persistence in a host associated niche. The predominance of annotations related to metabolism, catalytic activity, transport, cytochrome P450 families, and biosynthetic gene clusters is broadly consistent with the recognized metabolic plasticity of Aspergillus species [5, 21]. Transport related functions may reflect an enhanced capacity to accommodate extracellular stress [21–23], whereas diverse biosynthetic gene clusters may indicate broader ecological and chemical interaction potential [5, 6, 24]. Similarly, PHI-base and DFVF-linked matches support the presence of genes homologous to factors previously associated with pathogenicity related phenotypes [25, 26]. However, these results should be interpreted cautiously, because annotation reflects functional potential rather than direct evidence of expression or pathogenic activity in otomycosis [21]. Thus, the genomic dataset is best regarded as a pathogen contextual resource that refines mechanistic hypotheses rather than directly proving virulence or therapeutic response.

This genomic context also provides a rationale for exploring mechanistically distinct antifungal strategies. A. terreus AT-1 is clinically notable among non-fumigatus Aspergillus species because of its reported therapeutic complexity and reduced susceptibility to some antifungal agents in selected settings [27, 28]. Although the present study did not directly define the full susceptibility profile of this isolate to standard antifungal drugs, the genomic background strengthened the rationale for evaluating antimicrobial peptides as complementary candidates. In our study, genome wide annotation identified 291 putative resistance associated genes in the A. terreus AT-1 isolate, of which 34.7% (101/291) were membrane associated, and transporter candidates spanned other transport systems (48.6%, 34/70), ABC transporters (27.1%, 19/70), and MFS/PDR-type transporters (24.3%, 17/70). This transition is conceptually important because many antifungal peptides act primarily through membrane perturbation or other envelope directed effects rather than through a single intracellular target [7]. They are less prone to rapid target site mutation driven resistance. However, given the diversity of predicted transporter/envelope functions, tolerance/adaptation remains plausible and warrants monitoring during iterative peptide optimization. Therefore, such membrane active mechanisms are often considered advantageous because they may reduce the ease of rapid target specific resistance, although adaptive responses and tolerance remain possible [7, 28]. Accordingly, the genomic and peptide modules are linked not by direct causality, but by a therapeutic logic in which pathogen contextualization supports the testing of a mechanistically diverse intervention.

Within the peptide module, a notable finding was that constraining peptide length, net charge, and overall hydrophobic content did not eliminate meaningful biological divergence among the designed molecules. Although C1–C5 were engineered within a narrow physicochemical space, they still differed in antifungal activity and hemolytic behavior, with C3, C4, and C5 showing stronger activity against A. terreus AT-1, but C5 also displaying substantially greater hemolysis. This pattern suggests that bulk descriptors alone do not fully predict selectivity [8]. It is possible that residue arrangement, local amphiphilicity, and the distribution of aromatic side chains influenced peptide and membrane interactions in ways not captured by overall charge and hydrophobicity alone [8]. In addition, the F motif may have altered peptide self association or membrane insertion behavior, thereby affecting both potency and host cell compatibility [29]. These findings therefore underscore that short peptide optimization requires attention not only to global physicochemical parameters but also to sequence level organization. In this context, the superior performance of C3 may be attributed to a more favorable balance between cationic charge distribution, hydrophobic interactions, and phenylalanine-driven aromatic packing, which together likely enhance selective membrane partitioning and destabilization of fungal conidial membranes while maintaining reduced nonspecific interactions with host cells. This optimized amphiphilic arrangement may therefore account for the observed improvement in antifungal efficacy and biosafety profile relative to the other designed peptides.

C3 emerged as the most promising candidate because it preserved antifungal activity while remaining within an acceptable hemolytic range, and this advantage extended from vegetative cells to conidia. This is particularly relevant because residual conidia may contribute to persistence or relapse in otomycosis. Although the MIC of C3 against conidia was higher than that against vegetative cells, this difference is biologically plausible given the greater stress tolerance and protective surface architecture of conidia [22, 23, 30]. Importantly, both flow cytometric PI analysis and Hoechst/PI co-staining supported the same conclusion: C3 increased conidial membrane permeability, implicating membrane compromise as a key component of its antifungal action. While additional downstream effects cannot be excluded, these data provide a mechanistically interpretable cellular endpoint linking antifungal activity to membrane disruption [31].

Several limitations should be acknowledged. First, the clinical and genomic analyses were centered on a single otomycosis derived A. terreus AT-1 isolate, which limits generalizability across strains and settings [27, 28]. Second, the relationship between the genomic and peptide modules remained interpretive rather than causally resolved, and the mechanistic work focused mainly on membrane permeability. Nevertheless, these limitations do not negate the central conclusion that a genome-informed contextual approach can be productively paired with constrained short peptide design to identify candidates with an improved antifungal and biosafety balance.

Conclusion

Overall, this study provides a pathogen contextualized workflow for antifungal discovery in an otomycosis relevant setting. The genomic analyses refined the biological characterization of a clinical A. terreus AT-1 isolate, whereas the peptide analyses identified C3 as a lead candidate distinguished by antifungal activity, limited hemolysis, and membrane permeabilizing effects on conidia. These findings imply that short membrane active peptides may be valuable complementary antifungal candidates for localized fungal infections, particularly where therapeutic diversification beyond conventional agents is desirable [7, 8]. Future work should expand validation across larger isolate collections, compare peptide activity with standard antifungal agents and combination regimens, and further define the structural determinants that separate fungal selectivity from host cell toxicity.

Abbreviations

A. terreus

A. Aspergillus terreus

C. albicans

Candida albicans

C. krusei

Candida krusei

C. glabrata

Candida glabrata

ABC

ATP-binding cassette

MFS

Major facilitator superfamily

PDR

Pandrug resistance system

TRF

Tandem Repeats Finder

tRNA

Transfer RNA

rRNA

Ribosomal RNA

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

KOG

Clusters of Orthologous Groups

NR

Non-Redundant Protein Database

TCDB

Transporter Classification Database

CAZy database

The Carbohydrate-Active EnZymes database (CAZy)

PHI-base

Pathogen–Host Interactions database

DFVF

Database of Fungal Virulence Factors

BGC

biosynthetic gene cluster

NRPS

nonribosomal peptide synthetase

T1PKS

type I polyketide synthase

MIC

Minimum inhibitory concentration

ROS

reactive oxygen species

AMPs

Antimicrobial peptides

FCZ

fluconazole

PI

propidium iodide

Authors’ contributions

Conception and design: Wenguang Shi, Jiahao Hu, Qingru Jiang, Yingxiang Xu, Rongjie Li, Huihui Xu, Gaoyang Zhang, Zixuan Cui, Ruiyue Ouyang, Wenxuan Wu, Yujie Li, Haiyu Hong, Xiangyu Mou, Wenjing Zhao and Shuli ChouMethodology: Wenguang Shi, Jiahao Hu, Qingru Jiang, Yingxiang Xu, Rongjie Li, Huihui Xu, Gaoyang Zhang, Zixuan Cui, Ruiyue Ouyang, Wenxuan Wu, Yujie Li, Haiyu Hong, Xiangyu Mou, Wenjing Zhao and Shuli ChouAnalysis and interpretation of data: Wenguang Shi, Jiahao Hu, Qingru Jiang, Yingxiang Xu, Rongjie Li, Huihui Xu, Gaoyang Zhang, Zixuan Cui, Ruiyue Ouyang, Wenxuan Wu, Yujie Li, Haiyu Hong, Xiangyu Mou, Wenjing Zhao and Shuli ChouWriting, review, and/or revision of the manuscript: Wenguang Shi, Jiahao Hu, Qingru Jiang, Yingxiang Xu, Rongjie Li, Huihui Xu, Gaoyang Zhang, Zixuan Cui, Ruiyue Ouyang, Wenxuan Wu, Yujie Li, Haiyu Hong, Xiangyu Mou, Wenjing Zhao and Shuli ChouStudy supervision: Shuli Chou.

Funding

This work has been supported by the National Key Research and Development Program of China (Grant No. 2020YFA0907800), the Shenzhen Science and Technology Programs (KQTD20200820145822023, RCJC20231211085944057, and ZDSYS20220606100803007), and the National Natural Science Foundation of China (Grant No. 81671479).

Data availability

The genome sequence generated in this study has been deposited in the NCBI database under BioProject accession PRJNA1429839. The assembled genome is publicly available under accession number GCA_057814615.1, with associated BioSample accession SAMN56256256 and WGS accession JBYDZR000000000.

Declarations

Ethical approval and consent to participate

This study involving human participants and human-derived specimens was conducted in accordance with the Declaration of Helsinki. The cerumen samples used in this study were approved by the Ethics Committee of the Fifth Affiliated Hospital of Sun Yat-sen University (approval no. 2025-K30-1). The study was conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants. The samples were strictly limited to this study, and any remaining specimens were disposed of in accordance with institutional regulations after completion of the experiments. The red blood cells from male C57BL/6 mice (SPF grade, Zhuhai Biotest Biotechnology Co., Ltd, SCXK(GD) 2020-0051), aged 6-8 weeks and weighing approximately 25 g, were utilized in this experiment. The experimental protocol was approved by the Sun Yat-sen University Experimental Animal Ethics Committee, ensuring adherence to the principles of animal protection, animal welfare, and ethics (Approval No.: SYSU-IACUC-2024-003027). The protocol is also in compliance with the relevant national regulations on experimental animal welfare and ethics.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.